Rescue previously failed therapies
All pharmaceutical companies have many drugs that have failed in a previous clinical trial or were not approved by the FDA. Anywhere from $800 Million to $1.4 Billion and many years have been spent on each compound. What if a percentage of these drugs or their derivatives could be brought to market? Imagine the benefit to patients with novel therapies that previously failed due to underlying genetic variants rendering them susceptible to adverse events. Unlock value from the previous investments straight to the company bottom line.
GMLs, the solution for drug revitalization
Heligenics offers a new product to revitalize these failed drugs by stratifying genetically defined populations to enhance efficacy and safety, thus minimizing the main failure points for clinical trials. Through a massively parallel experimental test of drug’s effect upon cells, Heligenics uses the “GigaAssay” to produces Gene Mutation / Function Library (“GMLs”). Each GML experiment measures the impact of all possible amino acids substitutions in the drug target upon the targets molecular function. This experiment is repeated in the presence and absence of the drug.
To improve efficacy and safety, comparison of GMLs identifies substitutions that can produce resistance, including those in the drug binding site, and those causing cell toxicity. The variants in the GML’s are next compared to population allele frequencies to identify the likelihood that patients with a particular variant could adversely impact a clinical trial. Clinical trials are redesigned with the knowledge of major alleles of the drug target that can produce resistance or toxicity as defined by the GMLs. This knowledge is then used as exclusion criteria or arm stratification in a new trial, or used retrospectively from existing data if genetic data was collected during the original trial. Alternatively, knowledge of amino acid positions associated with resistance or toxicity can guide drug derivatization or new drug development to reduce these important unwanted effects, thereby reducing risk in the drug development and approval process. Heligenics will test your custom drugs and targets.
Many recently approved drugs (>140) have a genotype on the label, thus genetics is becoming increasingly important in drug testing. Heligenics GMLs will classify molecular function and cell toxicity of each variant, thereby inferring efficacy and safety for patients having any genetic variant in the drug target gene. Combined with allele frequency data and drug binding site data, clinical trials can be accurately designed.
Brivanib has had a massive investment and was tested for treatment of patients with hepatocellular cancer in 25 different clinical trials and has failed four Phase 3 trials, the last in 2017 (clinicaltrials.gov). In general the drug is well tolerated, but is failing trials for its poor efficacy, especially when compared to Sorafenib, the only approved drug on the market. The patient response the drug is variable with some participants having a positive response and others with no effect.
The drug targets are 7 receptors in the tyrosine kinase family, three Vascular Endothelial growth Factor Receptors 1-3 (VEGFR), and four Fibroblast Growth factor Receptors 1-4 (FEFR). One target, VEGFR2, is illustrated below, and is the target for Brvanib.
Protein Databank structure of VEGFR2 bound to AAL993 (PDB: 5EW3)
The structure for Brivanib complexed with a receptor was not available but there is a structure with another ATP analog, AAL993 that binds VEGFR2 (PDB: 5EW3). The contact residues for the drug with the binding site are: V848, A866, V867, K868, I892, E885, I888, L889, V898, V914, V916, E917, C919, L1019, L1035, I1044, D1046, F1047.
There are no mutations in VEGFR2 for hepatocellular cancer and only 17 variants for other disorders in ClinVar, the main public disease-variant database. However, there are 713 missense variants observed in ~250,000 people ranging from singletons to 22% allele frequency; 74 variants are in the drug binding site region. There are about 1,500 missense variants observed in cancers (COSMIC), with approximately 100 in this binding site region.
Shown above is a GML result for a fictitious gene; this table shows 5 variants from a table of over 10,000 SNVs (1). This type of GML has >97% of single nucleotide variants encoding single amino acid substitutions. For the Activity Metric (2), activity is measured against wild-type activity in the first column. A measurement of 100 is 1% of the wild-type (‘baseline’) activity. Pathogenicity (3) is determined from an activity threshold relative to wild-type activity and known pathogenic mutations.